|Publication number||US8001584 B2|
|Application number||US 11/241,589|
|Publication date||Aug 16, 2011|
|Filing date||Sep 30, 2005|
|Priority date||Sep 30, 2005|
|Also published as||CN1964258A, CN1964258B, EP1929745A1, EP1929745B1, US20070079362, WO2007041325A1|
|Publication number||11241589, 241589, US 8001584 B2, US 8001584B2, US-B2-8001584, US8001584 B2, US8001584B2|
|Inventors||Victor B. Lortz, Jesse Walker, Preston J. Hunt, Amol Kulkarni|
|Original Assignee||Intel Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (16), Non-Patent Citations (6), Referenced by (7), Classifications (13), Legal Events (2)|
|External Links: USPTO, USPTO Assignment, Espacenet|
The invention relates to network device configuration. More specifically, the invention relates to secure methods of configuring devices to gain access to network resources.
Wireless communication between computing devices has enjoyed wide adoption and significant growth as a flexible and cost-effective alternative to traditional hard-wired network infrastructure. Wireless technologies such as WiFi (a common name for several related standards proposed by the Institute of Electrical and Electronics Engineers, “IEEE”) and Bluetooth permit data transfer via radio signals in 2.4 GHz, 5 GHz, and other bands. New standards and improved equipment have increased data rates of wireless networks, but the technology has some issues that have not been satisfactorily addressed. Configurability and security of wireless networks are two of these.
Wireless networks rely on encryption of packets to prevent eavesdropping and unauthorized use of network resources. For example, the Wired Equivalent Policy (“WEP”), which is a part of IEEE standard 802.11 describing wireless communications, specifies the encryption to be used in WiFi networks. Likewise, Wi-Fi Protected Access (WPA) is an alternative encryption and authentication standard based on mechanisms defined in the IEEE 802.11i standard. However, products supporting WEP, WPA, and similar security standards typically are difficult to configure correctly, so wireless networks are often run in unencrypted, “open” mode. Furthermore, even when encryption is enabled on a wireless local area network (“WLAN”), the participating systems often lack a standardized way to change the security configuration. Easy-to-use, broadly-applicable procedures to configure and manage WLAN participants may be of considerable value in the field.
Embodiments of the invention are illustrated by way of example and not by way of limitation in the figures of the accompanying drawings in which like references indicate similar elements. It should be noted that references to “an” or “one” embodiment in this disclosure are not necessarily to the same embodiment, and such references mean “at least one.”
If the user of a device such as WLAN client laptop 170 wishes to use the wireless network through AP 110 to access resources on other wireless or wired nodes, he must obtain a valid encryption key and enter it into the wireless device's configuration. Traditionally, an administrator of the wireless network would provide the key and the user would type it into a configuration form. However, this approach is inconvenient for the user and cumbersome for the administrator. In addition, an unauthorized user may obtain a copy of the key from the user and use it to access the network. Changing the WLAN configuration to exclude such an unauthorized user may entail re-configuring all of the other authorized devices.
A superior method of managing access to the WLAN can be built on a registration protocol according to an embodiment of the invention. The protocol involves AP 110, new WLAN client 170 and a network entity called the registrar, shown in this figure as device 180. In other embodiments, the registrar may be integrated with the AP. Some networks may use several registrars.
Registrar 180 may communicate with AP 110 over the wired network 140, over a wireless (radio) connection, or both. The registrar may provide administrative facilities to monitor the WLAN and manage WEP encryption keys.
New WLAN client 170 has an associated identifier called a device password. The password may be engraved on the device or printed on a label, or may be displayed by the device or by software associated with the device. If the device password is displayed in this way, it may be dynamic (for example, the displayed password may be valid for a period of time or until some event occurs, then a new device password may be chosen and displayed). In some embodiments, the device password may be readable by a reader device near the new client. For example, Near Field Communication (“NFC”) devices can exchange data wirelessly over a short distance, so a device password might be stored in an NFC token and read by an NFC reader. In another embodiment, the new WLAN client might be equipped with an infrared or other light signal transmitter, and be able to transmit the device password to an optical receiver of the registrar within line-of-sight proximity.
Registrar 180, AP 110 and client 170 can interact according to the flow chart of
First, the client's device password is provided to the registrar (210). This may be accomplished by reading the password from the client's label or display and entering it through a registrar user interface, by placing the client near the registrar so that the registrar can read the client's NFC token automatically, or other similar method. This portion of the protocol may incorporate human involvement (for example, to read and enter the device password, or to place the client near the registrar). The remainder of the protocol can continue automatically.
Next, the client transmits a first message (“M1”) to initiate an EAP transaction (220). M1 contains a first random number N1 and a public key PKE of the client, and may contain other information (described below). M1 is received by the registrar (225).
The registrar responds to M1 by transmitting a second message (“M2”) containing a second random number N2 and a public key PKR of the registrar (230). The client receives M2. (235).
The transaction continues with the client transmitting a message Mn (240) and the registrar responding with message Mn+1 (250). Portions of each message may be encrypted with a key known to both the client and the registrar, or with a public or private key of one of the parties. Messages may have appended a keyed message authentication code (“HMAC”), containing a keyed cryptographic hash of the previous message and a portion of the current message preceding the HMAC, to permit the recipient to verify that the other party correctly received the previous message and that no third party is tampering with the messages in transit.
One or more of the messages from the registrar contain information that the client can use to verify that the registrar knows the client's device password. This permits the client to verify that it is receiving credentials from an authorized registrar (e.g. the registrar to which the device password was provided, and not, for example, a rogue registrar that is attempting to trick the client into connecting to a hostile wireless network). The message(s) need not contain the device password itself; instead, information from which the client can infer that the registrar must know the password is sufficient. This information can be thought of as a “proof-of-possession” of the device password.
One or more of the messages from the registrar contains a credential such as a WEP or WPA key that the client can use to access the wireless LAN through the AP. The credential may be encrypted with a credential key to prevent its recovery by an eavesdropper. When the client receives the message containing the credential, it verifies the proof-of-possession to ensure the message came from a registrar with knowledge of its own device password (260). If the proof fails or is unacceptable, the client aborts the EAP transaction by transmitting a negative acknowledge (“NACK”) message (265). If the proof-of-possession correctly demonstrates knowledge of the device password, the client may decrypt the credential and store it in a configuration database for future use (
Once the client has successfully received the credential, it transmits an EAP “Done” response to the registrar (280), which receives the “Done” message (285) and responds with an EAP “Fail” message (290). The client subsequently receives the “Fail” message (295). Note that in this context, the failure message does not mean that the client must repeat the EAP transaction to obtain a credential. It merely indicates that the transaction was used to provision a credential rather than to grant the client immediate use of the wireless LAN. The client may use the credential it received later, when it attempts to access the network through the AP (299). For example, the client may update its configuration according to data in the credential, or may use the credential to complete a new authentication protocol transaction designed to provide network access.
Table 1 shows the detailed structure of the eight (8) messages exchanged between client (“C”) and registrar (“R”) according to an embodiment of the invention.
Version || N1 || Description || PKE
Version|| N1 || N2 || Description || PKR
Version || N2 || E-Hash1 || E-Hash2 ||
Version || N1 || R-Hash1 || R-Hash2 ||
ENCKeyWrapKey(R-S1) || HMACAuthKey(M3||M4*)
Version || N2 || ENCKeyWrapKey(E-S1) ||
Version || N1 || ENCKeyWrapKey(R-S2) ||
Version || N2 || ENCKeyWrapKey(E-S2) ||
Version || N1 || ENCKeyWrapKey(Credential) ||
The following symbols and abbreviations are used in Table 1:
Concatenation of parameters
Message Mn (excluding a hash value suffix)
Protocol version number
128-bit random numbers
Text string describing a device that transmitted the
Diffie-Hellman public key of client
Diffie-Hellman public key of registrar
Two secret random numbers selected by client
Keyed cryptographic hashes of E-S1 and E-S2,
respectively (each hashed together with separate
halves of the client's device password)
Two secret random numbers selected by registrar
Keyed cryptographic hashes of R-S1 and R-S2,
respectively (each hashed together with separate
halves of the client's device password)
Item encrypted with Key
HMAC keyed hash of item using key Key
In the embodiment defined by messages M1-M8, the participants in the registration protocol identify themselves in their first messages (M1 and M2). The “Description” portion of those messages may provide a free-form description of the device, or may instead be a standardized text string. Some embodiments may include several identification substrings with contents, lengths and/or delimiters agreed upon by the parties participating in the protocol. If the protocol is implemented as an EAP method on top of the IEEE 802.1X framework, a predefined EAP Identity may be used to select that EAP method. For example, a client seeking a credential might use an EAP Identity string such as “SomePrefix-Enrollee-1-0”. In this standardized string, “1” and “0” may represent protocol or other version numbers; more advanced clients may transmit a similar string with higher integers instead of “1” and “0.” SomePrefix indicates a commonly-known string that the participants agree to use, and not the literal sequence of characters “SomePrefix.” An example commonly-known string is “WFA-SimpleConfig”.
As mentioned earlier, messages M3-M8 contain a message authentication code (“MAC”) to permit the recipient to verify that the protocol messages have not been corrupted or tampered with. In this embodiment, the MAC of a message is a cryptographic hash calculated over the data of the previous message and data of the current message, excluding the MAC portion of the current message. HMACkey is a keyed hash, which can only be generated or validated by a party that possesses the key. Selection of keys is discussed below.
In an embodiment that uses the eight messages shown in Table 1, note that the client and registrar each divide the device password into two portions and incrementally and Alternatingly prove knowledge of those two portions in several successive messages (M3-M7). This improves the security of the protocol by thwarting a potential attack to obtain the device password, as explained here with reference to
The client commits to prior knowledge of both halves of its device password by first transmitting E-Hash1 303 and E-Hash2 306 (310). E-Hash1 and E-Hash2 correspond to hashes of random numbers E-S1 and E-S2 with the first and second halves of the device password, respectively. For additional security, the keyed hash used to generate E-Hash1 303 and E-Hash2 306 may incorporate information to tie the hash values to the current enactment of the protocol. For example, the key of the keyed hash could depend in an irreversible way on the random numbers N1 and N2 in the first two messages.
The registrar likewise pre-commits to knowledge of both halves of the device password using R-Hash1 312 and R-Hash2 315 (320). The registrar also provides R-S1 318, which the client can use to recreate R-Hash1 (330). If the client finds that R-Hash1 does not match the hash value it calculates using the first portion of its true device password and R-S1 (335), it can abort the EAP transaction (by sending an EAP negative acknowledge (“NACK”)) (340), without disclosing any information about its true password. If the first portion of the device password from the registrar is verified, then the client transmits E-S1 in message M5 (350). E-S1 permits the registrar to verify that the previously-transmitted E-Hash1 303 was computed using the first portion of the same device password. The registrar verifies E-Hash1 with E-S1 (360) using the first portion of the device password it knows (
Several portions of messages M1-M8 are encrypted to prevent an eavesdropper from learning privileged information such as the device password or the credential. Some of the message parameters—for example, E-S1, E-S2, R-S1 and R-S2—may be random bit strings selected by either the client or the registrar. Other message parameters must be known to both entities so that one side can encrypt and/or authenticate a message and the other side can decrypt and/or authenticate it. Some embodiments of the invention will use a key derivation key (“KDK”) that is computed from the Diffie-Hellman secrets, random numbers N1 and N2, and a Media Access Control (“MAC”) address of the client. In one embodiment, the KDK is defined as:
where HMAC-SHA-256 is a keyed hash function using key DHKey, which is:
DHKey=SHA-256(gAB mod p)
PKE, the client public key, is gA mod p; and PKR, the registrar public key, is gB mod p. SHA-256 is a well-known secure hash algorithm and “mod” indicates modulo arithmetic. Alternate embodiments using other public key technologies such as DSA, RSA, or elliptic curve, may also be used. In these cases, M3 may include a proof-of-possession of the client's public key, and the registrar may encrypt the KDK using the client's public key and send it to the client in M4.
Additional keys may be derived from KDK using a key derivation function (“KDF”) to generate a plurality of pseudo-random bits according to an algorithm. A function that may be used by an embodiment to generate a key of n bits is described by the following pseudo-code:
KDF( Key, PersonalizationString, NBits ) :
Result = “”
Iterations = (NBits + PRFDigestSize − 1)/PRFDigestSize
for i = 1 to Iterations do
Result = Result +
Here, PRF is a function to produce pseudo-random key bits. Some embodiments may use a keyed hash such as HMAC-SHA-256 for this purpose. The AuthKey and KeyWrapKey keys used in messages M3-M8 may be derived as follows:
AuthKey∥KeyWrapKey∥ESMK=KDF(KDK, “Wi-Fi Easy and Secure Key Derivation”, 640)
AuthKey may be a 256-bit key used to authenticate messages M3-M8. KeyWrapKey may be a 128-bit key used to encrypt random keys E-S1, E-S2, R-S1 and R-S2 and the credential transferred in M8. ESMK may be a 256-bit key whose use will be discussed later. Summarizing the formulas and pseudo-code above, AuthKey, KeyWrapKey and ESMK are derived according to an algorithm that accepts as inputs the two random numbers N1 and N2, the client's MAC address, and a Diffie-Hellman secret. Subsets of the bits generated by KDF are selected and used for AuthKey, KeyWrapKey and ESMK.
In some embodiments, a portion of the protocol implemented through messages M1-M8 gcan be short-circuited. If a secure communication channel between client and registrar exists, the registrar can use that channel to transmit a credential for the client. For example, if the registrar and client can both use a removable storage medium such as a Universal Serial Bus (“USB”) Solid State Disk, then the registrar may write the credential in a file on the USB disk and the client may obtain the credential by reading the file. Information transmitted via the secure channel may still be encrypted to protect against unauthorized access and/or tampering, or to permit the client to verify that the credential came from an authorized registrar. For example, the credential may be encrypted with a symmetric encryption key derived from KDK. In another embodiment, the credential may be encrypted using the client's public RSA key. The fact that the credential is delivered across a trusted out-of-band channel serves to verify the credential's authenticity.
The protocol described above can be used in several additional situations. First, consider the problem of associating a new registrar with an existing access point (“AP”). The protocol can operate between the new registrar and the AP, with the AP taking the role of the client. This use of the protocol might be indicated by a different EAP Identity string. For example, the string “SomePrefix-Registrar-1-0” could indicate to the AP that a registrar wished to associate itself with the AP. As with the previous client identity string, the numbers “1” and “0” might indicate a protocol version, and could be replaced with larger integers by a more advanced registrar. Some protocol messages may be modified to carry information of use in this scenario. For example, the AP may include information about its present configuration when it transmits M7. The configuration information may be encrypted with a configuration key such as KeyWrapKey, which is already used to encrypt E-S2 in that message. The registrar, upon receiving the AP's present configuration, may prepare an updated or new configuration and transmit it to the AP as part of the credential in message M8. The new configuration would also be encrypted in that message.
The protocol could be used again, after a client device had successfully received a credential, if new credentials were to be distributed. This use of the protocol is known as “rekeying.” A client participating in a rekeying operation might use a different value for the device password. In one embodiment, the device password could be a 256-bit pseudo-random bit string generated by the Key Derivation Function described earlier:
The foregoing descriptions and explanations detail a secure protocol by which two entities can authenticate each other and transfer a credential over an insecure network such as an 802.11 wireless network or other radio network, a public access network, or the like. The protocol may be used even when the client does not yet have a credential that the AP will accept. In the WLAN case, the client may structure its messages (and receive its replies) according to the IEEE 802.1X protocol. An AP may accept 802.1X-formatted messages and forward them to a registrar (or process them internally, if the AP itself contains a registrar), even though the client transmitting the messages lacks an acceptable WEP key, credential or other security arrangement for secure communication.
Messages according to an embodiment of the invention may also be transmitted over a wired or wireless network as data encapsulated in a higher level protocol such as the User Datagram Protocol (“UDP”) or Transmission Control Protocol (“TCP”), running over the Internet Protocol (“IP”). Alternatively, messages could be formatted in the Extensible Markup Language (“XML”) and embedded in Hypertext Transfer Protocol (“HTTP”) transactions according to the Universal Plug-n-Play (“UPnP™”) standard promulgated by the UPnP Forum.
An embodiment of the invention may be a machine-readable medium having stored thereon instructions which cause a processor to perform operations as described above. In other embodiments, the operations might be performed by specific hardware components that contain hardwired logic. Those operations might alternatively be performed by any combination of programmed computer components and custom hardware components.
A machine-readable medium may include any mechanism for storing or transmitting information in a form readable by a machine (e.g., a computer), including but not limited to Compact Disc Read-Only Memory (CD-ROMs), Read-Only Memory (ROMs), Random Access Memory (RAM), Erasable Programmable Read-Only Memory (EPROM), and a transmission over the Internet.
The applications of the present invention have been described largely by reference to specific examples and in terms of particular allocations of functionality to certain hardware and/or software components. However, those of skill in the art will recognize that credentials can be exchanged over public and/or insecure networks by devices consisting of software and hardware that distribute the functions of embodiments of this invention differently than herein described. Such variations and implementations are understood to be apprehended according to the following claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US7454619 *||Sep 5, 2003||Nov 18, 2008||Palo Alto Research Center Incorporated||Method, apparatus, and program product for securely presenting situation information|
|US20030115453||Dec 17, 2001||Jun 19, 2003||Grawrock David W.||Connecting a virtual token to a physical token|
|US20040019786 *||Oct 14, 2002||Jan 29, 2004||Zorn Glen W.||Lightweight extensible authentication protocol password preprocessing|
|US20040111600 *||Dec 5, 2002||Jun 10, 2004||Kaler Christopher G.||Deriving keys used to securely process electronic messages|
|US20040158738 *||Jan 23, 2004||Aug 12, 2004||Fujitsu Limited||Security management device and security management method|
|US20050135625 *||Oct 27, 2004||Jun 23, 2005||Yoshimichi Tanizawa||Communication apparatus and method|
|US20050177733||Aug 16, 2002||Aug 11, 2005||Togewa Holding Ag||Method and system for gsm authentication during wlan roaming|
|US20060156388 *||Jan 13, 2005||Jul 13, 2006||Vlad Stirbu||Method and apparatus for a security framework that enables identity and access control services|
|US20060259759 *||May 16, 2005||Nov 16, 2006||Fabio Maino||Method and apparatus for securely extending a protected network through secure intermediation of AAA information|
|US20070005955 *||Jun 29, 2005||Jan 4, 2007||Microsoft Corporation||Establishing secure mutual trust using an insecure password|
|US20070127719 *||Oct 13, 2004||Jun 7, 2007||Goran Selander||Efficient management of cryptographic key generations|
|US20070288743 *||Aug 22, 2007||Dec 13, 2007||Cisco Technology, Inc.||Enabling stateless server-based pre-shared secrets|
|EP1513288A2||Sep 1, 2004||Mar 9, 2005||Xerox Corporation||Method, apparatus, and program product for provisioning secure wireless sensors|
|EP1538779A1||Oct 14, 2003||Jun 8, 2005||Matsushita Electric Industrial Co., Ltd.||Identification information protection method in wlan interconnection|
|WO2004028071A1||Nov 25, 2002||Apr 1, 2004||Nokia Corp||Linked authentication protocols|
|WO2005067685A2||Jan 10, 2005||Jul 28, 2005||Cisco Tech Ind||Enabling stateless server-based pre-shared secrets|
|1||International Application No. PCT/US2006/038114 Int'l Search Report & Written Opinion dated Feb. 19, 2007.|
|2||Int'l Preliminary Report on Patentability (Chapter 1) issued Apr. 1, 2008 for International Application No. PCT/US2006/038114.|
|3||Office Action for Chinese Patent Application No. 200610163914.5 mailed Aug. 31, 2010, 7 pgs.|
|4||Office Action for Chinese Patent Application No. 200610163914.5 mailed Dec. 25, 2009, 22 pgs.|
|5||Office Action for European Patent Application No. 06 804 267.0 mailed Sep. 9, 2010, 10 pgs.|
|6||Office Action mailed Apr. 24, 2009 for Chinese Patent Application No. 200610163914.5 with translation, Whole document.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US8504836 *||Dec 29, 2008||Aug 6, 2013||Motorola Mobility Llc||Secure and efficient domain key distribution for device registration|
|US8769627 *||Dec 8, 2011||Jul 1, 2014||Symantec Corporation||Systems and methods for validating ownership of deduplicated data|
|US8904172||Jun 16, 2010||Dec 2, 2014||Motorola Mobility Llc||Communicating a device descriptor between two devices when registering onto a network|
|US8950000 *||Nov 15, 2012||Feb 3, 2015||Sprint Communications Company L.P.||Application digital rights management (DRM) and portability using a mobile device for authentication|
|US9100382||Mar 15, 2013||Aug 4, 2015||Qualcomm Incorporated||Network security configuration using short-range wireless communication|
|US9148423||Dec 29, 2008||Sep 29, 2015||Google Technology Holdings LLC||Personal identification number (PIN) generation between two devices in a network|
|US20100169646 *||Dec 29, 2008||Jul 1, 2010||General Instrument Corporation||Secure and efficient domain key distribution for device registration|
|U.S. Classification||726/5, 713/168, 713/156|
|Cooperative Classification||H04L2209/56, H04L9/3226, H04L9/3234, H04L2209/805, H04L63/162, H04L63/0807|
|European Classification||H04L9/32, H04L63/16B, H04L63/08A|
|Jan 13, 2006||AS||Assignment|
Owner name: INTEL CORPORATION, CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:LORTZ, VICTOR B.;WALKER, JESSE;HUNT, PRESTON J.;AND OTHERS;REEL/FRAME:017013/0723;SIGNING DATES FROM 20051214 TO 20060112
Owner name: INTEL CORPORATION, CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:LORTZ, VICTOR B.;WALKER, JESSE;HUNT, PRESTON J.;AND OTHERS;SIGNING DATES FROM 20051214 TO 20060112;REEL/FRAME:017013/0723
|Feb 4, 2015||FPAY||Fee payment|
Year of fee payment: 4